2-December Offline Report Matthias Schröder Topics: Monte Carlo Production New Linux Version Tape Handling Desktop Computers

2-December Monte Carlo Production Csaba Hajdu has taken over the responsibility for production system He has made himself familiar with the system Ran a few test runs Now awaiting your request for Monte Carlo production Once the OPAL environment is ready for SLC3, he will check compatibility

2-December New Linux version New linux version (SLC3) is now certified –Common effort between CERN and Fermilab Hepix scripts have been fixed by for SLC3 is the same as for redhat 7.3 –Add a special directory ‘i386slc3’ to paths /afs/cern.ch/opal/offline/geant/pro/i386slc3/lib /afs/cern.ch/opal/offline/gopal/pro/i386slc3/lib /afs/cern.ch/opal/offline/rope/pro/i386slc3/lib /afs/cern.ch/opal/offline/mgr//i386slc3/ –Have to set path variables accordingly –Can hopefully be supressed in next linux version… /bin/link hides /afs/cern.ch/opal/offline/scripts/link –Add alias for ‘link’ to OPAL environment

2-December New Linux, cont. OPAL tools are being prepared –Scripts need updating hardware fort Link ? ROPE code should be ok –Old executables can be used on new system But not the other way around –Mixing of libraries: ???? But in any case I prefer to produce separate libs for the new system MC production: Csaba will check as soon as the environment is ready Lxplus and lxbatch will by and by be migrated to SLC3 –The lxplus alias will point to lxslc3 on –Use lxplus7 to get to lxplus running RedHat 7.3! Already available now –RedHat 7.3 capacity will be kept until Easter 2005? Most likely at least until summer…

2-December Tape Issues Castor files that are registered in fatmen by tape & FSN had been migrated –IT did not warn us –Fatmen pointed to wrong tape -> I/O errors –Have updated ~ 290k fatmen entries in summer –More to come tricky cases of the ranges done in summer (3664 entries) Other media types –In general fatmen still ‘primary’ catalog Not all Monte Carlo runs are in castor Have stopped TMS usage –TMS: Tape Management System Obsolete, hard to maintain Planned to switch off early this year Kept running due to OPAL’s manpower situation –Got rid of calls within ROPE (easy, done), checked Monte Carlo Production system –Switched off service as a test, no problems observed, finally switched off some weeks ago

2-December OPAL Desktop Computers The OPAL clusters (lxopal, hpopal) don’t exist any more All desktop machines fully independent You can shutdown your machine or move it whenever you need to Move: have to update netops database and IP number CERN machine: better ask me first… Institutes must nominate a ‘Responsible’ for their machines To be registered in netops database The responsible is also responsible for installing kernel updates and security in general Machines for which we don’t get a responsible will be disconnected from network and trashed Bologna, Freiburg, Heidelberg, QMC, Riverside still have ‘Opal Linux’ as responsible… Machines can be administered by user or responsible CERN machines: responsible is me Password will be given to responsible Help for general issues can be provided via the Helpdesk Help for OPAL issues from the usual suspects Make sure your data on local disks is archived

2-December HPUX Computers Shall we trash them, or do you want to keep one as souvenir?

2-December Conclusion OPAL offline quiet –Except for the busy periods… –Essential systems kept running with minimal effort –Difficult to free resources (time!) for major changes Many thanks to Akos for keeping the Monte Carlo Production going Many thanks to Csaba for taking over the Monte Carlo Production Make sure your data on local disks is archived